Sex as an Independent Variable in the Measurement of Satiation: A Retrospective Cohort Study

Abstract

Background:

Satiation is a key component of food intake regulation as it brings an eating episode to an end. The effect of sex on satiation measurement has not been characterized.

Objective:

To assess the effects of biological variables on satiation.

Design:

Retrospective cohort study. We included 959 participants (mean age 39 [SD 12] years; 70.7% female, and BMI 33 kg/m² [8]) who had measurements of satiation with a nutrient-drink test to assess volume to fullness (VTF) and maximum tolerated volume (MTV), and/or an ad libitum meal test to assess calories consumed to fullness (CTF). We performed univariate and multiple regression analyses to estimate the contribution of sex to VTF, MTV, and CTF, compared to other biological variables, such as age, weight, height, BMI, waist-to-hip circumference (W/H), and lean mass percentage (LM%), that are known to affect these parameters.

Results:

Females had higher BMI, W/H, and LM%. VTF, MTV, and CTF were lower in females: 704 [323] vs. 783 [328] mL, p=0.001; 1226 [384] vs. 1419 [410] mL, p<0.001; and 871 [291] vs. 1086 [326] kcal, p<0.001; respectively. Sex was a strong and independent predictor of VTF, MTF and CTF: parameter estimate [PE]=−80.8, p=0.006; PE=−124.2, p=0.0007; and PE=−110, p=0.001; respectively.

Conclusions:

Sex has a strong effect on satiation measured by VTF, MTV, and CTF, even after adjusting for other biological factors known to affect these parameters. Females seem to integrate intra-meal inhibition signals to consume fewer calories unrelated to body size or composition.

INTRODUCTION

The interaction between the gastrointestinal and central nervous systems is essential for food intake regulation. Satiation is one of the key manifestations of appetite control, as it reflects the constellation of physiologic processes that promote meal termination, thereby limiting meal size and energy intake(1). Satiation is regulated by neurohormonal signals secreted by the gut in response to food’s macronutrient and mechanical properties(2). Decreased satiation is associated with a larger meal size, contributing to increased caloric intake and the development of overweight and obesity(3).

The importance of sex as a variable in overweight and obesity is illustrated by the prevalence of obesity worldwide, with about twice as many females as males suffering from severe obesity(4). Pre-clinical and clinical studies have revealed female-specific factors in the two physiological determinants of obesity: energy intake and energy expenditure(5–10). Data to date have revealed that certain foods have higher satiating effects in females compared to males(11–16). Although the mechanisms behind this phenomenon are not fully understood, estrogens seem to play an essential role in the central regulation of satiation(5).

Despite these observations, there is limited data on how sex affects indices of satiation. Several approaches are available to measure satiation through ad libitum intake of food or drinks while monitoring calories consumed before fullness sensation is reached(17, 18). One of these approaches is the nutrient drink test, a validated tool that determines volumes associated with meal termination called the volume to fullness (VTF) and the maximum tolerated volume (MTV)(19, 20). Another validated approach is the ad libitum meal test, in which the calories of food consumed to fullness (CTF) are recorded (CTF)(18). The nutrient drink test allows for the understanding of the continuum of fullness sensation until maximal fullness is reached with liquid calories at predetermined intervals. The ad libitum meal test accounts for daily caloric intake variations by measuring calories to fullness with a substantial meal consumed after a standardized breakfast. Measurements of satiation are determined by factors that can be classified as physiologic (e.g., fasting gastric volume or capacity and gastric emptying), behavioral (e.g., eating patterns and behaviors), food-dependent (e.g., texture, presentation, and palatability), and environmental(18, 21). Demographic and anthropometric variables, such as sex, age, and body mass index (BMI) have been shown to affect some of the aforementioned factors in health and disease (19, 22–24). Despite the complexity of satiation regulation, current reports on this topic are limited by the small sample size and the number of covariates used to investigate sex differences in measurements of satiation.

For this study, we hypothesized that satiation measured by VTF and MTV during the nutrient drink test and CTF at an ad libitum meal test, two standardized satiation tests, differs between females and males after adjusting for biological variables, some with clear sexual dimorphism, and that could theoretically affect measurements of satiation. This study aimed to assess the effect of sex compared to other biological variables such as age, weight, height, BMI, waist-to-hip circumference (W/H), and lean mass percentage (LM%) on satiation parameters among otherwise healthy participants across all BMI classes.

SUBJECTS AND METHODS

The Institutional Review Board approved this retrospective cohort study. We collected data on 959 unique participants from previous prospective studies or those currently participating in active trials in our research unit(3, 25–27). We collected baseline data (i.e., before any intervention) in participants with stable weight (i.e., less than ±3 kg weight change within three months of recruitment). We included participants across all BMI classes and with controlled weight-associated comorbidities. We excluded participants with a history of eating or substance abuse disorders and those taking medications reported to alter gastric motility or bodyweight, such as prokinetics or glucagon-like peptide 1 receptor agonists/analogs. Females with childbearing potential were required to have a negative pregnancy test 24 hours before enrollment. All participants signed informed consent before enrollment in the original studies, and they consented to the use of their data, medical records, and samples for research (Supplemental Figure 1).

Measurement of Satiation

All the patients included had undergone tests to measure satiation, including the nutrient drink test and/or the ad libitum meal test. We minimized all external food cues in our research setting during satiation tests. Participants had no time or drink/food limit, and food allergies or restrictions were assessed prior to any test.

• Nutrient drink test: After an overnight 8-hour fast, we measured satiation by determining the VTF and the MTV. We assessed VTF and MTV using a validated nutrient drink test(19). For this test, patients ingested Ensure^® (1 kcal/ml, 11% fat, 73% carbohydrate, and 16% protein) at a constant rate of 30 ml/min. We used a numerical scale graded from 0 to 5 to record the progression of their fullness sensation at 5-minute intervals throughout the test (table 1). Patients were instructed to drink as much as they wanted until reaching level 5 of fullness. After reaching level 5, the test was stopped, and VTF and MTV were calculated from the volumes of nutrient drink ingested when participants reached scores of 3 and 5, respectively(19).

• Ad libitum meal test: We measured satiation by determining the CTF during the ad libitum meal test. The ad libitum meal was performed 4 hours after a 320kcal standard breakfast consisting of eggs, toast, and skim milk. The ad libitum meal test consisted of all-you-can-eat vegetable lasagna [Stouffers^®, Nestle USA, Inc, Solon, OH; nutritional analysis of each 326g box: 420kcal, 17g protein (16% of energy), 38g carbohydrate (37% of energy), and 22g fat (47% of energy)]; vanilla pudding [Hunts^®, Kraft Foods North America, Tarrytown, NY; nutritional analysis of each 99g carton: 130kcal, 1g protein (3% of energy), 21g carbohydrate (65% of energy), and 4.5g fat (32% of energy)]; and skim milk [236mL carton: 90kcal, 8g protein (36% of energy), 13g carbohydrate (64% of energy), and 0g fat]. Participants were instructed to eat as much as they wanted until reaching maximal fullness. All food and drink were presented identically on each occasion and covertly weighed before and after the meal. After reaching maximal fullness, energy and macronutrient intakes were calculated using nutritional values provided by the manufacturer and by a registered dietitian, using validated software (ProNutra 3.0; Viocare Technologies Inc, Princeton, NJ).

Table 1.

Numerical Scale to Determine Fullness Progression

Numerical scale	Fullness sensation
0	No symptoms
1	First sensation of fullness
2
3	Volume to fullness, i.e., fullness following a typical meal a
4
5	Maximum tolerated volume b

^aNumerical value used to assess satiation

^bNumerical value used to evaluate nature of dyspeptic symptoms

Body composition

Height and weight were assessed after 8-hour fasting at arrival to the research unit. Body composition (fat and lean mass) was measured by dual-energy X-ray absorptiometry (General Electric (GE) Lunar iDXA).

Outcomes

The primary outcome of this analysis was to compare VTF and MTV by nutrient drink test and CTF during an ad libitum meal test in females and males, considering other biological variables that can potentially affect satiation, including age, height, weight, W/H, and LM%.

Statistical Analysis

Statistical analysis was performed using JMP (Statistics software v JMO Pro14.1.0; SAS Institute Inc., Cary, NC). All normally distributed continuous data are summarized as means ± standard deviations (SD) unless otherwise indicated. Categorical data are presented as frequencies and percentages. We used Pearson χ² and two-sided t-test to make between-group comparisons for nominal and ordinal variables. For continuous variables, we used Wilcoxon rank-sum test. We used Pearson linear (R) correlation coefficient to measure associations. We performed linear regression to study the effect between covariates. To estimate the contributions of biological and anthropometric variables to VTF and MTV in the nutrient drink test and CTF in the ad libitum meal test, we conducted a multiple regression analysis with and without interactions with the “sex” variable and summarized the results based on parameter estimates (PE) and significance values. All p values <0.05 were considered statistically significant.

RESULTS

Demographics and anthropometric data

Our cohort included 959 participants. Most participants were female (678 [70.7%]) and Caucasian (835 [87.1%]) with a mean age of 38.6±12.1 years (Table 2). The mean BMI was 33.5±7.7 kg/m². Most of the participants had obesity (646 [67.3%]), followed by overweight and normal weight (191 [19.9%] and 122 [12.7%], respectively). Females had a higher BMI compared to males: 33.9 vs. 32.8 kg/m², respectively (p=0.04). This difference was driven by a higher proportion of female patients with obesity compared to males: 70% vs. 62% (p=0.01). There were significant differences between females and males in height, weight, waist and hip measurements, and body composition (Table 2).

Table 2.

Demographic and Anthropometric Characteristics of the Cohort, and Primary Outcomes (n=959)

A. Demographic and Anthropometric Characteristics of the Cohort
		All (n = 959)	Female (n = 678, 70.7%)	Male (n = 281, 29.3%)	P value
Age, years		38.6 ± 12.1	39.1 ± 11.7	37.4 ± 12.9	0.01
Race, Caucasian		835 (87.1%)	603 (88.9%)	232 (82.6%)	0.60
Height, m		1.7 ± 0.08	1.6 ± 0.07	1.8 ± 0.09	<0.001
Weight, kg		96.6 ± 24.2	94.5 ± 22.7	101.7 ± 26.8	<0.001
Waist circumference, cm		115.8 ± 14.6	114.1 ± 13.9	121.6 ± 15.7	<0.001
Hip circumference, cm		126.2 ± 13.1	127.2 ± 12.9	122.9 ± 12.9	0.009
W/H		0.91 ± 0.07	0.89 ± 0.07	0.98 ± 0.06	<0.001
BMI, kg/m2		33.5 ± 7.7	33.9 ± 7.7	32.8 ± 7.5	0.04
BMI Class	Normal weight, n	122 (12.7%)	88 (13.0%)	35 (12.1%)	0.60
	Overweight, n	191 (19.9%)	117 (17.3%)	74 (26.3%)	0.001
	Obesity, n	646 (67.3%)	473 (69.8%)	173 (61.6%)	0.01
Body Composition	Total Fat Mass, kg	52.8 ± 15.9	52.7 ± 13.4	52.9 ± 18.5	0.95
	Fat Mass Percentage, %	46.4 ± 6.5	50.5 ± 4.6	42.3 ± 5.2	0.001
	Total Lean Mass, kg	59.8 ± 8.4	50.4 ± 6.3	69.2 ± 10.5	0.001
	Lean Mass Percentage, %	51.0 ± 6.0	49.3 ± 4.7	56.6± 6.2	<0.0001
B. Primary Outcomes (n=959)
Nutrient Drink Test (n = 556)	VTF, mL	729.6 ± 326.6	703.9 ± 323.14	782.9 ± 328.3	0.001
	MD ± SED: 79.0±29.6, 95% CI 21.9 to 139.2
	MTV, mL	1288.2 ± 402.5	1222.6 ± 386.2	1419.5 ± 403.5	<0.0001
	MD±SED: 196.9±34.6, 95% CI 132.9 to 269.2
Ad libitum meal test (n=631)	CTF, Kcal	927.9 ± 315.5	871.2 ± 292.1	1085.6 ± 325.6	<0.0001
	MD±SED: 214.4±28.6, 95% CI 160.9 to 273.9

Continuous data are summarized as mean ± standard deviation. Categorical data are presented as frequencies and percentages.

All P-values <0.05 were considered significant.

Abbreviations used: BMI, body mass index; CI, Confidence interval; CTF, calories to fullness; Kcal, Kilocalories; mL, MD, Mean difference, Milliliter; Min, minutes; MTV, maximum tolerated volume; SED, Standard error difference; VTF, volume to fullness; W/H, waist-to-hip ratio.

Measurements of Satiation

Five hundred and fifty-six participants (57.6%) underwent the nutrient drink test, and 631 (65.3%) underwent the ad libitum meal test. Two hundred and fifty-nine participants (26.8%) underwent both tests.

The average VTF and MTV for the cohort were 729.6±326.6mL and 1288.2±402.5mL, respectively. VTF and MTV were significantly lower in females compared to males: 703.9±323.14 vs. 782.9±328.3 mL, p=0.001 [Mean difference (MD) ± Standard error difference (SED) 79.0±29.6 mL, 95% Confidence Interval (CI) 21.9 to 139.2 mL] and 1222.6±386.2 vs. 1419.5±403.5mL, p<0.0001 [MD±SED 196.9±34.6 mL, 95% CI 132.9 to 269.2 mL]; respectively (Table 2B and Figures 1A and B). In the ad libitum meal test, the mean CTF consumed was 927.9±315.5 kcal. Energy intake at the ad libitum meal test was significantly lower in females compared to males: 871.2±292.1 vs. 1085.6±325.6 kcal, p<.0001 [MD±SED 214.4±28.6 kcal, 95% CI 160.9 to 273.9 kcal] (Table 2B and Figure 1C). There were positive and strong correlations between VTF and CTF (R=0.42, p<0.0001) and between MTV and CTF (R=0.57, p<0.0001). Linear regression demonstrated that the intercepts for both analyses, were significant (PE ± Standard error [SE]: 298.9±60.2, p<0.0001 and 511±72, p<0.0001, respectively) (Supplemental Figure 2).

Figure 1.

Volume to Fulness and Maximum Tolerated Volume by Nutrient Drink Test, and Calories to Fullness by Ad Libitum Meal, by Sex.

Univariate Analysis and Linear Regression

As shown in table 3, VTF had a positive and significant correlation with weight (R=0.19, p<0.0001), height (R=0.16, p=0.0002), and BMI (R=0.22, p=0.004). MTV had a positive and significant correlation with weight (R=0.18, p<0.0001) and height (R=0.30, p<0.0001). CTF had a positive and significant correlation with weight (R=0.20, p<0.0001), height (R=0.40, p<0.0001), LM% (R=0.21, p=0.001), and W/H (R=0.24, p<0.0001). In our cohort, age did not correlate with VTF, MTV, or CTF. Interaction analysis between sex and BMI as a continuous variable and between sex and BMI class (normal weight, overweight, and obesity) for each of the primary outcomes (VTF, MTV, and CTF), and we found no significant interaction between the variables.

Table 3.

Predictors of Satiation Measurements: Univariate Analysis. Correlations Between VTF, MTV, and CTF, and Demographics and Anthropometric Variables

	VTF			MTV			CTF
	R	P	n	R	P	n	R	P	n
Age	−0.04	ns	556	0.07	ns	559	−0.06	ns	631
Weight	0.19	<0.0001	546	0.18	<0.0001	550	0.20	<0.0001	622
Height	0.16	0.0002	547	0.30	<0.0001	551	0.40	<0.0001	623
BMI	0.22	0.0042	549	0.04	ns	587	0.02	ns	625
LM %	a	a	a	a	a	a	0.21	0.001	246
W/H	a	a	a	a	a	a	0.24	<0.0001	366

All P-values <0.05 were considered significant.

^aData not available

Abbreviations used: BMI, body mass index; CTF, calories to fullness; LM%, lean mass percentage; MTV, maximum tolerated volume; VTF, volume to fullness; W/H, waist-to-hip ratio.

Multiple Regression Analysis

Table 4 shows the multiple regression analysis. For VTF, sex and BMI were both strong and independent predictors, with male sex and higher BMI predicting a higher VTF. For MTV, sex and height were both strong and independent predictors, with male sex and higher height predicting a higher MTV. For CTF, only sex was an independent predictor of calories consumed during an ad libitum meal test, with the female sex predicting a lower CTF. We performed sex variable interaction with all the covariates in each model and sex does not affect the relation between other covariates and VTF, MTV, and CTF with one exception. In the multiple regression analysis between CTF, female sex, and the other covariates, lean mass percentage was a predictor of CTF only in females and not in males (PE −11,2, 95% CI −18.8; −3.6, p=0.004).

Table 4.

Multiple regression Analysis

A. VTF Adjusted for Age, Sex, Height, and Weight
	Parameter estimate	95% Confidence Interval	P value
Age, older/younger	−1.1	−3.2; 0.1	0.3
Sex, female/male	−80.8	−138.8; −22.8	0.006
BMI, higher/lower	6.7	2.3; 11.1	0.003
Intercept ± SE	581.8 ± 77.7		<0.0001
B. MTV Adjusted for Age, Sex, Height, and Weight
	Parameter estimate	95% Confidence Interval	P value
Age, older/younger	3.0	−0.1; 5.5	0.06
Sex, female/male	−124.2	−196.8; −52.8	0.0007
Height, higher/lower	1022.3	627.7; 1417	<0.0001
Weight, higher/lower	1.4	−0.4; 3.2	0.14
Intercept ± SE	−651.0 ± 327.0		0.04
C. CTF Adjusted for Age, Sex, Height, and Weight
	Parameter estimate	95% Confidence Interval	P value
Age, older/younger	−0.6	−4.2; 7.7	0.7
Sex, female/male	−221.6	−356.6; −86.6	0.001
Height, taller/shorter	549.0	−16.0; 1114.0	0.06
W/H, larger/smaller	234.0	−331.7; 799.7	0.4
LM%, higher/lower	0.6	−6.4; 7.7	0.8
Intercept ± SE	−214.6 ± 618.8		0.73

All P-values <0.05 were considered significant.

Abbreviations used: BMI, Body Mass Index; CTF: Calories to Fullness; LM%, Lean Mass Percentage; MTV, Maximum Tolerated Volume; SE, Standard Error; VTF, Volume to Fullness; W/H, Waist-to-Hip Ratio.

DISCUSSION

This is the largest and most comprehensive cohort of participants with normal weight, overweight, and obesity, characterizing sex differences in satiation, measured by VTF and MTV utilizing a nutrient drink test, and CTF measured by the ad libitum meal test. It is worth noting that multiple reports have established the reproducibility of the nutrient drink test and ad libitum meal test (using a preload paradigm as we did in this study), with good intrasubject correlation and no significant differences between measurements at different visits(28–32). We showed that satiation measurements with both methods correlate significantly. There is high variance between measures that can potentially be related to the systematic difference between measures that is corroborated by the significant intercept in the simple linear regressions: the nutrient drink test is performed after an 8-hour fasting (no preload caloric intake) compared to the ad libitum meal test which is performed 4 hours after the consumption of a 320-cal standardized breakfast. Importantly, we showed that VTF and MTV by nutrient drink test, and CTF by an ad libitum meal test, all validated indices of satiation, differ between female and male participants across all BMI classes, with female sex having lower VTF, MTV, and CTF. On multiple regression analysis, we showed that sex was the only independent predictor for all three indices. These differences were not explained by other biological variables such as age, weight, height, BMI, W/H, or LM%, variables that could affect satiation(22, 33, 34).

Our study demonstrated that females reached satiation at lower volumes and/or calories, consistent with other smaller studies(19, 22, 23). Delgado-Aros et al. showed that significantly fewer calories (i.e., lower volumes) were required to reach maximum fullness for females compared to males in a cohort of healthy participants (n=134) across a wide age range (12–60 years old) and across all BMI classes(22). In this report, sex differences observed in VTF were independent of BMI, weight, or height. No other covariates, such lean mass percentage or waist-to-hip ratio were taken into consideration. Similarly, a study in patients with obesity only demonstrated that male sex was a predictor of higher VTF in the nutrient drink test, although the study only included five males among 62 participants(25). This same study reported that sex was a significant predictor of CTF by an ad libitum meal test but failed to perform multiple regression analysis to investigate whether sex differences were the results of differences in other parameters such as BMI or gastric volumes. It is paradoxical that women perceive satiation at lower volumes and/or calories compared to men and still have a greater prevalence of obesity. In general, women have lower energy requirements than men due to differences in lean mass content. We did not have data on body composition to see whether sex differences observed in VTF and MTV were independent of LM%. However, we showed that when adjusted for LM%, sex was an independent predictor for CTF measured by the ad libitum meal test only in female but not male participants. As reported by others, age was not an important determinant of satiation in this cohort of adults(22). It is important to note that some studies have found that patients with obesity have a greater gastric capacity that could result in greater caloric/volume intake before gastric distention-stimulated satiation occurs(22, 33, 34). Because we included all BMI classes, from normal weight to obesity, and based on the interaction analysis between sex and BMI, we demonstrated that the sex differences observed were independent of BMI or BMI class.

The disproportional prevalence and susceptibility to obesity in females, despite the increased satiation perceptions, suggests that sex-specific metabolic processes might drive the difference in indices of satiation. To corroborate the importance of sex on food intake regulation, Gonzalez-Izundegui et al. demonstrated that gastric emptying was slower in females than in males, and slower gastric emptying was associated with decreased appetite and increased fullness, and lower caloric intake(27). Additionally, neuroimaging studies have demonstrated different brain responses to ingestive stimuli between females and males despite both groups reporting similar appetite sensations(10). This evidence suggests a sex-specific activation pattern in brain areas that might predispose females to specific food cues, particularly those in the occipital area (i.e., visual). Yunker et. al. demonstrated that female patients with obesity have a higher BOLD-fMRI activity to food cues in the orbitofrontal and prefrontal cortices after sucralose consumption(35). Finally, data show that sex hormones play a critical role in food intake regulation. Sex hormones independently influence food intake as observed by the cyclical fluctuations observed in females during the follicular and luteal phases of the menstrual cycle, with an increased intake during the latter(6). These findings, taken together, demonstrate the critical contribution of sex in regulating the gut-brain axis through gastric emptying and appetite sensations but do not explain the higher prevalence and greater severity of obesity in females. The sexual dimorphism of obesity might be teleological with an evolutionary basis(36).

Energy intake is tightly regulated by the homeostatic and hedonic components of food intake that depend on conscious and unconscious neurohormonal processes(37). On one side, the homeostatic component of food intake regulation is driven by satiation, satiety, and hunger. While satiation is the result of intra-prandial physiologic processes that promote meal termination, satiety is the result of post-prandial physiologic events that determine the timing for the next meal by inhibiting the sensation of hunger. Satiation, satiety, and hunger are controlled by signals from the gut, the vagus nerve, and the adipose tissue. On the other hand, the hedonic component of food intake regulation is characterized by the drive to eat solely to elicit pleasure, disregarding the nutritional value of food and/or the individual’s metabolic status. There is a complex crosstalk between the homeostatic and hedonic components of food intake with the latter one having the capacity to override the former. Differences in other components of the homeostatic control of food intake (i.e., satiety and hunger), hedonic eating, and/or the crosstalk mechanisms between these components may drive the differences in the prevalence and severity of obesity. For instance, data support gender-differences in hunger, food cravings, and food-cue reactivity, controlled by the homeostatic or hedonic components of food intake, or a crosstalk between both, that may explain, at least in part, gender differences observed in obesity’s epidemiologic and clinical trends (10, 14, 38–42). However, most studies have been done in small cohorts of healthy individuals, mostly without overweight or obesity, and were characterized by female overrepresentation. Another factor that may participate in the differences observed in gastric capacity. Data have revealed that fasting gastric volume correlates with VTF (25). This relation has not been established with calorie consumption during an ad libitum meal test. Interestingly, post postprandial changes in gastric volume (or gastric accommodation) do not correlate with VTF or MTV, or CTF(25). We do not have data on fasting gastric volume and/or accommodation for our cohort to assess its effect on sex differences in satiation.

Obesity is a heterogeneous disease. Physiologic studies focused on identifying specific, actionable phenotypes are contributing to understanding the pathophysiologic processes underlying obesity and, hence, the development of individualized strategies for better weight loss outcomes(3, 43). Quantitative and qualitative measurements to assess food intake regulation are hard to achieve, mainly due to the many systems involved, the complexity of their connections, and the effect of important biological variables on these systems and their connections. Among the biological variables, anthropometric measures, particularly weight and height, have been historically considered the most important factors regulating food intake. Most recently, sex has been demonstrated to be a very important variable, one that regulates mechanisms controlling food intake independently of other anthropometric factors. Although males are on average taller and heavier than females, and one could infer that this is the reason why women eat less than men, we have demonstrated that sex, even after correction for anthropometric measures, is an independent predictor of satiation, one of the key mechanisms regulating food intake. The mechanisms behind this remain unknown. In the meantime, as women are disproportionately affected by obesity and sex is a fundamental source of variation in physiology, sex should be systematically considered when studying the pathophysiology of obesity and its therapeutic interventions (7, 11, 12, 15, 16, 44). Unveiling the sex-specific neurohormonal mechanisms in food intake regulation is essential to understanding the difference in obesity prevalence and severity in females and will aid in the identification of therapeutic targets to enhance weight loss outcomes and reduce weight-associated comorbidities. For instance, if variability in ovarian hormones affects the hedonic and homeostatic components of food intake regulation, or any other process that regulates energy balance, hormonal replacement therapy could be used to minimize the effect on weight changes.

Our study has several strengths. Most importantly, this study is the largest to date to characterize VTF and MTV by nutrient drink test and CTF by ad libitum meal test in adult females and males in a cohort of participants across all BMI classes. Also, as satiation perception can be influenced by many parameters (e.g., presentation, palatability, and texture of foods) that can elicit different sensory responses, we controlled for these possible confounders by rigorously standardizing the drink/food stimuli, the research setting, and potential external food cues.

Our study has some limitations. First and most importantly, this was a retrospective observational analysis, we could not investigate the mechanisms behind sex differences in satiation. Second, women’s energy intake fluctuates within the menstrual cycle with distinct periods of increased caloric intake. We do not have any data on the menstrual cycle and satiation test timing to assess if female sexual hormones explain, at least partially, the differences seen. We also do not have data on menopausal status, information that could help elucidate if ovarian hormones directly affect satiation measurement. More research is needed to investigate the role of the hormonal effect of sex on satiation. Third, to assess fullness sensation during the nutrient drink test, we used a 5-point scale. Appetite sensations such as fullness, satisfaction, and appetite can also be measured with linear scales between 100 and 150 mm. Although one could argue that linear scales are better suited to measure appetite sensations as they allow a finer grade of discrimination in patients’ responses, these scales are not optimal for the nutrient drink test. Linear scales can help us measure MTV which would coincide with the extreme of the scale. However, linear scales do not allow for the measurement of VTF as no studies have been completed to identify the point or range in the linear scale that correlates with VTF. Fourth, during the ad libitum meal test, we asked patients to eat until reaching maximal fullness which limits the possibility of calculating the rate at which satiation is achieved. However, because it is a complete meal that contains all the major food groups, it is possible to model more real-world scenarios and concentrate on the caloric content rather than the timing at which satiation is reached. Finally, satiation can be strongly influenced by subjective food preferences, and the limited options with our standard ad libitum meal and drink tests might have influenced the degree of energy intake.

CONCLUSION

In summary, this study highlights the effects of sex on satiation measured by VTF and MTV in the nutrient drink test and CTF in the ad libitum meal test. Sex is a strong independent predictor of satiation. These observations highlight the importance of sex as a variable in understanding the pathophysiologic mechanisms involved in the heterogeneity of obesity. This observation could help develop sex-specific strategies to enhance weight loss outcomes.

Abbreviations:

BMI

Body mass index

CTF

calories consumed to fullness

LM%

lean mass percentage

SD

standard deviations

VTF

volume to fullness

W/H

waist-to-hip circumference

DATA SHARING

Data collected for the study, including individual de-identified participant data, as well as study protocol, and informed consent will be available to interested parties with publication, after signing of a data access agreement. Data may be requested by contacting Dr. Andres Acosta M.D, Ph.D., at Acosta.andres@mayo.edu.